<p><strong>MiniMax H3 Prompt Generator (Portable)</strong></p><p>A local desktop tool that generates properly formatted prompts for MiniMax Hailuo 3.0 (H3) video. Supports Reference-to-video + audio (RF2VA), Text-only, First frame, and First+Last frame modes. Runs completely offline after setup using your own Qwen GGUF model.</p><h3 id="features">Features</h3><ul><li><p>Dark UI with image drag-and-drop</p></li><li><p>Live token streaming and cancel button</p></li><li><p>Strict H3 template output</p></li><li><p>Fully portable — no ComfyUI or system Python required</p></li><li><p>Auto-opens the browser</p></li></ul><h3 id="requirements">Requirements</h3><ul><li><p>Windows</p></li><li><p>NVIDIA GPU with a recent driver (CUDA 13.x compatible recommended)</p></li><li><p>A Qwen3.8-VL GGUF + matching mmproj file</p></li></ul><h3 id="setup">Setup</h3><ol><li><p>Download and unzip the package.</p></li><li><p>Put your GGUF and mmproj files inside the models folder (or set the paths in the UI).</p></li><li><p>Double-click run.bat.</p></li><li><p>The browser opens automatically to <a target="_blank" rel="ugc" href="http://localhost:7860">http://localhost:7860</a>.</p></li></ol><h3 id="notes">Notes</h3><ul><li><p>First load can take a minute while the model loads into VRAM.</p></li><li><p>The tool does not include the model — you supply your own.</p></li><li><p>GPU offload works if the CUDA wheel and runtime are present.</p></li></ul><p></p>

Files with SHA256: 9b5e12d98c53c9b5637792c84d4f207435902d6b3d893e1f6fc7d5aa8a07646c